Cross-Modal Learning of Housing Quality in Amsterdam (2403.08915v1)
Abstract: In our research we test data and models for the recognition of housing quality in the city of Amsterdam from ground-level and aerial imagery. For ground-level images we compare Google StreetView (GSV) to Flickr images. Our results show that GSV predicts the most accurate building quality scores, approximately 30% better than using only aerial images. However, we find that through careful filtering and by using the right pre-trained model, Flickr image features combined with aerial image features are able to halve the performance gap to GSV features from 30% to 15%. Our results indicate that there are viable alternatives to GSV for liveability factor prediction, which is encouraging as GSV images are more difficult to acquire and not always available.
- United Nations, Department of Economic and Social Affairs, Population Division, “World Urbanization Prospects: The 2018 Revision,” New York, Tech. Rep. ST/ESA/SER.A/420, 2019.
- G. W. Evans, “The built environment and mental health,” J Urban Health, vol. 80, no. 4, pp. 536–555, Dec. 2003.
- S. Thompson and J. Kent, “Healthy Built Environments Supporting Everyday Occupations: Current Thinking in Urban Planning,” Journal of Occupational Science, vol. 21, no. 1, pp. 25–41, Jan. 2014.
- S. Barber, D. A. Hickson, I. Kawachi, S. V. Subramanian, and F. Earls, “Neighborhood Disadvantage and Cumulative Biological Risk Among a Socioeconomically Diverse Sample of African American Adults: An Examination in the Jackson Heart Study,” J Racial Ethn Health Disparities, vol. 3, no. 3, pp. 444–456, 2016.
- A. Dubey, N. Naik, D. Parikh, R. Raskar, and C. A. Hidalgo, “Deep Learning the City: Quantifying Urban Perception at a Global Scale,” in ECCV 2016, B. Leibe, J. Matas, N. Sebe, and M. Welling, Eds. Cham: Springer, 2016, pp. 196–212.
- N. Naik, J. Philipoom, R. Raskar, and C. Hidalgo, “Streetscore – Predicting the Perceived Safety of One Million Streetscapes,” in CVPR 2014. Columbus, OH, USA: IEEE, Jun. 2014, pp. 793–799. [Online]. Available: http://ieeexplore.ieee.org/lpdocs/epic03/wrapper.htm?arnumber=6910072
- S. Scepanovic, S. Joglekar, S. Law, and D. Quercia, “Jane Jacobs in the Sky: Predicting Urban Vitality with Open Satellite Data,” ACM Human-Computer Interaction, vol. 5, pp. 1–25, Apr. 2021.
- A. Levering, D. Marcos, and D. Tuia, “Liveability from Above: Understanding Quality of Life with Overhead Imagery and Deep Neural Networks,” in Proceedings of IGARSS 2021. Brussels: IEEE, Jul. 2021.
- PDOK, “NIEUW: hogere resolutie luchtfoto als open data bij PDOK,” Oct. 2017. [Online]. Available: https://www.pdok.nl/-/nieuw-hogere-resolutie-luchtfoto-als-open-data-bij-pdok
- S. Srivastava, J. E. Vargas-Muñoz, and D. Tuia, “Understanding urban landuse from the above and ground perspectives: a deep learning, multimodal solution,” Remote Sensing of Environment, vol. 228, pp. 129–143, Jul. 2019, arXiv: 1905.01752. [Online]. Available: http://arxiv.org/abs/1905.01752
- I. Havinga, P. W. Bogaart, L. Hein, and D. Tuia, “Defining and spatially modelling cultural ecosystem services using crowdsourced data,” Ecosystem Services, vol. 43, p. 101091, Jun. 2020. [Online]. Available: https://www.sciencedirect.com/science/article/pii/S2212041620300334
- K. He, X. Zhang, S. Ren, and J. Sun, “Deep Residual Learning for Image Recognition,” in CVPR 2016, Jun. 2016, pp. 770–778.
- S. Srivastava, J. E. Vargas Muñoz, S. Lobry, and D. Tuia, “Fine-grained landuse characterization using ground-based pictures: a deep learning solution based on globally available data,” IJGIS, vol. 34, no. 6, pp. 1117–1136, Jun. 2020. [Online]. Available: https://doi.org/10.1080/13658816.2018.1542698
- M. G. Kendall, “A New Measure for Rank Correlation,” Biometrika, vol. 30, no. 1-2, pp. 81–93, Jun. 1938. [Online]. Available: https://doi.org/10.1093/biomet/30.1-2.81
- B. Zhou, A. Lapedriza, A. Khosla, A. Oliva, and A. Torralba, “Places: A 10 Million Image Database for Scene Recognition,” IEEE TPAM, vol. 40, no. 6, pp. 1452–1464, Jun. 2018.
Paper Prompts
Sign up for free to create and run prompts on this paper using GPT-5.
Top Community Prompts
Collections
Sign up for free to add this paper to one or more collections.